09-004 UCB Gen IV Base Isolation Rpt Final

Final Report Pg. 1 of 132

ADVANCED SEISMIC BASE ISOLATION METHODS FOR MODULAR

REACTORS

E. Blandford, E. Keldrauk, M. Laufer, M. Mieler, J. Wei,

B. Stojadinovic, and P.F. Peterson

Departments of Civil and Environmental Engineering and Nuclear Engineering

University of California

Berkeley, California

September 30, 2009

UCBTH-09-004

FINAL REPORT

Abstract

Advanced technologies for structural design and construction have the potential for

major impact not only on nuclear power plant construction time and cost, but also on the

design process and on the safety, security and reliability of next generation of nuclear

power plants. In future Generation IV (Gen IV) reactors, structural and seismic design

should be much more closely integrated with the design of nuclear and industrial safety

systems, physical security systems, and international safeguards systems. Overall

reliability will be increased, through the use of replaceable and modular equipment, and

through design to facilitate on-line monitoring, in-service inspection, maintenance,

replacement, and decommissioning. Economics will also receive high design priority,

through integrated engineering efforts to optimize building arrangements to minimize

building heights and footprints. Finally, the licensing approach will be transformed by

becoming increasingly performance based and technology neutral, using best-estimate

simulation methods with uncertainty and margin quantification.

In this context, two structural engineering technologies, seismic base isolation and

modular steel-plate/concrete composite structural walls, are investigated. These

technologies have major potential to (1) enable standardized reactor designs to be

deployed across a wider range of sites, (2) reduce the impact of uncertainties related to

site-specific seismic conditions, and (3) alleviate reactor equipment qualification

requirements. For Gen IV reactors the potential for deliberate crashes of large aircraft

must also be considered in design. This report concludes that base-isolated structures

should be decoupled from the reactor external event exclusion system. As an example, a

scoping analysis is performed for a rectangular, decoupled external event shell designed

as a grillage. This report also reviews modular construction technology, particularly

steel-plate/concrete construction using factory prefabricated structural modules, for

application to external event shell and base isolated structures.


CONTENTS

1.0 Introduction ...................................................................................................................................... 4 References ........................................................................................................................................................................ 7

2.0 Base Isolation, External Event, and Modular Construction Design Options .............. 8 2.1 Seismic Base Isolation ....................................................................................................................................... 8

2.1.1 Dynamics of a Seismically Isolated Structure ........................................................................ 9 2.1.2. Types of Horizontal and Vertical Seismic Isolation Devices ......................................... 12 2.1.3. Supplemental Damping and the Isolation Gap .................................................................... 14 2.1.3. Seismic Isolation Design Basis ................................................................................................... 14 2.1.4 Benefits and Challenges of Seismic Isolation ........................................................................ 16

2.2 External Event Shielding ................................................................................................................................ 17 2.3.1 Coupled above-grade event shell and citadel ...................................................................... 18 2.3.2 Decoupled above-grade event shell and citadel ................................................................. 20 2.3.3 Berms, below grade, and underground construction ...................................................... 23

2.3 Steel-Plate/Concrete Modular Construction ........................................................................................ 24 2.4 References ............................................................................................................................................................. 29

3.0 Simulation Verification and Validation Issues .................................................................. 31 3.1 Historical Experience with Seismic Base Isolation ............................................................................ 31 3.2 Current Analysis Methods and Tools for Reactor Seismic Design ............................................... 34 3.3 Verification and Validation Requirements and Approach ............................................................. 35 3.4 References ............................................................................................................................................................. 36

4.0 Loading Characterization .......................................................................................................... 38 4.1 Earthquake ............................................................................................................................................................ 38 4.2 Aircraft crash loading ...................................................................................................................................... 40 4.3 Blast loading ........................................................................................................................................................ 44 4.4 Internal and other loads ................................................................................................................................ 47 4.5 References ............................................................................................................................................................. 48

5.0 Integration With Reactor Systems ......................................................................................... 50 5.1 Technology-Neutral Framework (TNF) for Reactor Licensing ................................................... 50

5.1.1 Event Identification and Reliability Functions ................................................................... 51 5.1.2 Physical Arrangement and Reactor Safety ........................................................................... 52 5.1.3 Physical Arrangement and Physical Security Functions................................................. 54 5.1.4 Major SSCs Requiring Design Integration ............................................................................. 55

5.2 Reactor System Seismic Design Criteria .................................................................................................. 55 5.2.1 SSC Seismic Classification ............................................................................................................ 56 5.2.2 Implications for Plant Construction ......................................................................................... 57 5.2.2 Modularization .................................................................................................................................. 58

5.3 References ............................................................................................................................................................. 59

6.0 Parametric Analysis .................................................................................................................... 60 6.1 Base-isolated NPP Model for Seismic Analysis: Assumptions and Parameters ..................... 60 6.3 Earthquake Response of the Base Isolated NPP Model ..................................................................... 62 6.4 Response of Base-Isolated NPP Model to Aircraft Impact ............................................................... 67 6.5 Decoupled Event Shell Response to Aircraft Crash ............................................................................. 69 6.5 References ............................................................................................................................................................. 76

7.0 Findings and Conclusions .......................................................................................................... 77


8.0 Nomenclature ................................................................................................................................ 80

Appendix A: Ground motion ............................................................................................................ 81

Appendix B: Earthquake Ground Motion .................................................................................... 83

Appendix C: Earthquake Response Results .............................................................................. 113

Appendix D: Impact Response Time Histories ........................................................................ 115


1.0 INTRODUCTION

Structural engineering is commonly pursued at the tail end of the development of new

reactor designs, after the major nuclear and mechanical systems have already been

designed. As illustrated in Fig. 1, consideration of major structure functional

requirements, such as aircraft crash resistance, late in the design process can result in

expensive modifications. For Generation IV reactor options, nuclear, mechanical and

structural engineering should be integrated in design from the start to optimize the design,

increase its safety and security, and improve its economy.

Fig. 1. External event shell steel grillage surrounds the Lucas Heights reactor in

Australia.

A recent review of advanced construction technologies for nuclear power plants

(Schlaseman, 2004) identified twelve new structural systems and construction and

fabrication methods that have a high potential to improve the economy and reduce time

required to build new nuclear power plants. The common theme among these

technologies is modularization, prefabrication and preassembly. However, the report

failed to identify seismic base isolation, a technology already deployed to protect

conventional building structures and infrastructure facilities from earthquakes, as the

paradigm-changing structural design concept that enables modular construction.

Seismic base isolation has the potential to provide significant benefits by simplifying

nuclear power plant design requirements. Base-isolation devices limit the amount of

seismic energy transferred from the ground to the structures and filter out the high-

frequency components of horizontal ground motion: this makes it possible to design

standard, site-independent, nuclear power plant structures, systems and components

above the base isolation layer. Such designs will be inherently modular. Furthermore,

control of the magnitude and frequency content of the seismic input energy enables

design and implementation of new reactor concepts. It also helps toward standardization

since the same equipment and components can be re-used in multiple sites. For example,

seismic base isolation is considered essential for the deployment of pool-type Sodium


Fast Reactors (SFRs) and Advanced High Temperature Reactors (ATHRs), where the

reduced horizontal accelerations enable the use of thin-walled reactor vessels. Chapter 2

presents the fundamental principles of seismic base isolation and reviews current base

isolator technology.

The U.S. Advanced Liquid Metal Reactor (ALMR) program developed the first fully

engineered seismic base isolation system using (at the time, the 1980s) novel laminated rubber and steel sheet isolators. A summary of the analytical studies on the base isolation

of the prototype PRISM and SAFR designs as well as experimental studies to assess the

lateral deformation capacity and vertical stiffness of layered rubber isolators is presented

in by Taijirian et al. (1990). The 1994 ALMR design (ALMR, 1994) used 66 seismic

isolators to support a total base-isolated structure weight of 20,900 t (46,000 kips), which

included the reactor primary system, steam generators, and associated building structures.

The ALMR power conversion system, which consists of a steam turbine power plant fed

by two independent reactor modules, was not base isolated. The isolators stiffness was selected to provide a fundamental vibration period of 1.33 seconds, thus filtering out

seismic input energy associated with shorter vibration periods (higher frequencies).

While the PRISM plant has not been built, seismic base isolation has been used in power

reactors constructed at the Koeberg site in South Africa.

The most important new structural design criterion for new U.S. nuclear power

plants, since the original development of the ALMR design, is a new U.S. Nuclear

Regulatory Commission (USNRC) requirement that all new plant designs seeking design

certification include analysis for crash of a commercial aircraft into the reactor building.

Popular discussion of reactor structures commonly refers to a containment that is capable of protecting internal equipment from severe external events and containing

radioactive materials that might be released in a core-damage accident. In modern

reactor structures these two functions are separated.

In this report, we refer to the external event shell as the external walls of a reactor

building, which accommodates and transfers loads from external missiles such as aircraft,

and prevents penetration of objects or fuel into the building. Most commonly, due to

their substantial height (typical LWRs and SFRs are 75 to 80 m tall, while AHTRs are 35

to 40 m tall), the reactor building is constructed partially below grade. However, the

reactor building can also be constructed fully below grade, with berms, or deep

underground at the expense of increased excavation and retaining wall costs, and

complications for equipment and personnel access. The Humboldt Bay nuclear plant, a

65-MWe boiling water reactor built in 1963 in California, used below grade construction

where a caisson was sunk and plugged and the plant built inside it. But underground

construction has not been used subsequently for large nuclear plants. Inside the external

event shell, there is a citadel that provides the containment or confinement function for

retaining radioactive material that might be released from a reactor core during an

accident.

If an airplane crashes into a seismically base isolated structure, the crash will impart

momentum to the base isolated structure as a whole, with peak linear and angular

accelerations being strongly dependent upon the isolated building mass and on the

aircraft mass, speed, and location and direction of impact. In general, for lighter modular


reactors the imparted acceleration may be excessive, even if the mass of the power

conversion system is added to the base isolated mass. Alternatively, if the external events

shell is coupled to the ground, or if the structure is located below grade, limited or no

loads will be transmitted into the base isolated structure. Aircraft crash into a reactor

building will also generate large localized forces, induce local inelastic response and

damage, and produce debris that may act as interior missiles. New methods for steel-

plate/concrete modular construction, now being used in the Westinghouse AP-1000

reactor, have substantially improved local inelastic response of the external event shell

(called a shield building by Westinghouse) and reduced the chances for penetration and

generation of interior missiles, compared to conventional reinforced concrete

construction. Nevertheless, sufficient space should be provided between the external

event shell and the citadel so that localized inelastic response of the shell or differential

movement between the shell and the citadel does not result in damage to the citadel.

Chapter 2 discusses the different configuration and construction options for the external

event shielding and the citadel.

In order to use simulation results for base-isolated nuclear structures in a USNRC

Design Certification application, a significant amount of effort can be expected to

demonstrate that the results accurately represent the anticipated physical system. Chapter

3 reviews some of the general verification and validation challenges for the modeling and

simulation of base-isolated structures. The discussion includes observations and insights

based on limited historical experiences, currently accepted simulation methods, and the

representation of the non-linear behavior of base-isolators under large deformations. An

approach for the validation of models for isolated structures is presented with anticipated

experimental needs.

Reactor-building structural loads arise from several sources. Chapter 4 reviews the

loads that are created by earthquakes, aircraft crashes, blasts, and internal loads generated

by reactor equipment under normal and accident conditions.

Chapter 5 discusses the role of integrating key reactor safety systems with both

seismic base isolation and event shell design options. The design certification process for

Gen IV reactor types is currently in the speculative phase with the regulators focus being primarily on ALWR build proposals. However, it is reasonable to expect that technology-

neutral framework will emerge for Gen IV reactors that will be performance-based and

risk-informed with some additional conservatism due to potential issues commonly

associated with first-of-a-kind technology. The design certification process provides the

reactor vendor with the basis for design optimization while ensuring required operational

performance. An additional focus of Chapter 4 is to investigate the role of transforming

this process into a technology-neutral one where new structural design options can be

compared and further evaluated. The role of reactor design criteria with regard to safety

and security is also discussed.

Chapter 6 presents simplified parametric analysis for modular reactor structural

performance to earthquake and airplane impact excitations. The approach is consistent

with ASCE 43-05 principles (ASCE, 2005), which are under study but have not as yet

been endorsed by the NRC. Four performance issues are addressed. First, building lateral

and vertical displacements as well as rotations due to accidental torsion and off-center


impact are investigated to evaluate the ability of the isolation system to sustain them.

Second, potential for non-linear response of the structure above the isolation systems

level are investigated. Third, the interface with and impact on reactor safety functions

(reactivity control, decay heat removal, containment) and on balance-of-plant (BOP)

interface design is evaluated for different safety system and BOP design options. This

evaluation provides data allowing a quantitative comparison of the considered base

isolation design options.

There is very strong feedback between the design approaches taken to achieve

seismic base isolation, to accommodate new requirements for aircraft crash resistance, to

accelerate construction using steel-plate/concrete structural modules, to achieve physical

security and access control, to control reactivity and decay heat removal, and the design

of the balance of plant systems including power conversion. These choices impact

design, regulatory review and permitting, as well as construction materials quantities, and

construction time, and ultimately affect construction cost. Chapter 7 reviews these

issues, to provide a basis for making informed choices between these design options early

in the development of Gen IV modular reactors.

References

ALMR Reactor Facility Seismic Analysis, Bechtel National, Inc., October 1994.

American Society of Civil Engineers, Seismic Design Criteria for Structures, Systems and Components in Nuclear Facilities, ASCE Standard No. 43-05, p. 103, 2005.

Schlaseman, C. Application of Advanced Construction Technologies to Nuclear Power Plants, Technical Report MPR-2610, Revision 2, p. 132, US DoE, September 24, 2009.

Tajirian, F.F., J.M. Kelley and I.D. Aiken, Seismic Isolation for Advanced Nuclear Power Stations, Earthquake Spectra, Vol. 6, No. 2, pp. 371401, 1990.


2.0 BASE ISOLATION, EXTERNAL EVENT, AND MODULAR

CONSTRUCTION DESIGN OPTIONS

Multiple technical options exist for implementing seismic base isolation, external

event shielding, and modular construction. This chapter reviews these options and

discusses tradeoffs.

2.1 Seismic Base Isolation

Structural engineers typically employ one of two strategies to protect buildings from

the damaging effects of earthquakes. In the first strategy, engineers design the lateral

force resisting system to be very stiff, producing a structure with a short natural period

(or high natural frequency), small drifts and large accelerations in design earthquake

shaking. Such framing systems will protect drift-sensitive components (such as masonry

walls) but could damage acceleration-sensitive components (including mechanical

equipment, computers, and electrical systems). Historically, engineers of nuclear

structures have used this design strategy to make sure that horizontal deformations in

nuclear power plants are very small. Reinforced concrete nuclear power plant structures

have been designed using stiff structural systems such as reinforced concrete shear walls

or steel braced frames that remain essentially elastic in design earthquake shaking but are

detailed per ACI 349 (ACI 2006) to respond in a ductile manner in the event of shaking

more intense than the design basis. Nuclear plant structures also commonly use very

thick concrete structures to provide radiation shielding. These approaches differ from

conventional building design wherein substantial inelastic response and damage is

anticipated in design basis shaking.

In the second strategy, engineers design the lateral force resisting system to be

flexible, producing a structure with a long natural period (or small natural frequency).

Moment frames of structural steel and reinforced concrete are typical flexible seismic

framing systems. As a result, earthquake-induced horizontal accelerations and forces are

(relatively) small, but transient drifts are large, requiring structural elements and details

that allow for large deformation capacity and significant resilience, and residual drifts

might be significant. This design approach is common for steam turbine buildings,

particularly for pressurized water reactor (PWR) plants. The turbine-generator may be

spring mounted to isolate it vertically and horizontally. Gen IV plants may use compact

closed gas cycles, which would likely be mounted on the same base-isolated foundation

as the reactor. Vertical isolation may still be required for turbine-generator equipment.

Seismic isolation is a structural design approach that aims to control the response of a

structure to horizontal ground motion through the installation of a horizontally flexible

and vertically stiff layer of structural isolation hardware between the superstructure and

its substructure. The dynamics of the structure is thus changed such that the fundamental

vibration period of the isolated structural system is significantly longer than that of the

original, non-isolated structure, leading to a significant reduction in the accelerations and

forces transmitted to the isolated structure and significant displacements in the

deformable, structural base isolation, layer. The structural base isolation devices are

designed to sustain such large deformations without damage and to return the structure to

its original positions, thus leaving no residual deformation. The structure above the


isolation layer is designed to sustain the (reduced, and well-known) horizontal forces

transmitted through the isolation layer, usually such that it responds in its elastic response

range and remains undamaged. Vertical forces are transferred un-attenuated, but in

general it is easier to design structures and equipment to sustain vertical forces, and

where necessary, individual equipment items can be isolated for vertical forces using

spring-damper systems. The design of the isolator system must include sufficient space

for access to inspect and replace individual isolators. The components of a horizontally

seismically isolated structure are shown in Figure 2.

Fig. 2. Components of a base isolated structure shown in elevation.

Application of base isolation spread to the nuclear power industry with the completed

construction of plants in Cruas, France (1983) and Koeberg, South Africa (1984). Both

plants utilized neoprene pads and sliders, a system later deemed inappropriate for seismic

application in the United States and subsequently rejected in future plans in favor of

newer isolation types (Malushte and Whittaker, 2005). To date, there are 6 nuclear power

plants utilizing base isolation, all in France or South Africa.

Despite the lack of approved base isolated NPP designs in Japan and the United

states, regulations exist in both countries which could eventually make them a reality.

High seismicity and the expectation of high-frequency ground motions have led to harsh

design guidelines which retard the approval of NPP designs in the US. US Nuclear

Regulatory Commission Regulatory Guide 1.165 (NRC, 1997) imposes the strict

requirement that the seismic design of NPPs be based on a probabilistic seismic hazard assessment (PSHA) for a 100,000-year return period. Although many contend that base

isolation is a cost-effective way to achieve adequate response under such extreme loading

events, the relative infancy of the technology and limited in-field data of its response

hinder its implementation in new NPP designs.

2.1.1 Dynamics of a Seismically Isolated Structure

The seismic design acceleration and displacement spectra shown in Fig. 3 illustrate

the trade-off between horizontal displacements and horizontal accelerations in

seismically isolated structures. A seismic design spectrum plots the peak spectral

acceleration or spectral displacement versus the fundamental vibration period of a

structure under expected earthquake ground motion excitation. Fundamental vibration


periods for an isolated and fixed-base (or non-isolated) structure are shown: clearly, as

the fundamental vibration period of the structure increases (the structure becomes more

flexible), horizontal accelerations and, therefore, the seismic inertial forces decrease.

Simultaneously, horizontal displacements increase as the vibration period increases.

These increased displacements of the isolated superstructure, however, are concentrated

in the seismic isolation layer because it is much more flexible than the superstructure.

These displacements are accommodated by providing enough space around the

superstructure, a seismic isolation gap (Fig. 2), for it to freely travel during an

earthquake. The deformation (inter-story drifts) and horizontal floor accelerations in the

isolated superstructure will be smaller than those in its fixed-base counterpart because the

inertial forces acting on the superstructure are much smaller.

The efficacy of any isolation system is dependent on the ability of the isolators to

alter the fundamental period of the structure such that it is significantly larger than that of

the non-isolated structure, inducing a response that is far past the acceleration-sensitive

region of the earthquake spectrum (Chopra, 2007). This means that base isolation is most

appropriate for structures with naturally short non-isolated periods, such as nuclear

reactor buildings, existing in environments where damaging earthquake motions are

expected to have short predominant excitation periods. In such environments, the period

shift will greatly reduce superstructure accelerations and drifts. Care must be exercised,

however, to consider potential earthquake motions that may include significant longer-

period accelerations, as may be the case for softer types of soil conditions. An example is

presented in Chapter 6. A typical force-displacement relationship for a seismic isolation

device or system is shown in Figure 4. The isolation device is expected to respond to an

increase in displacement in a non-linear manner: the yield point (designated as (Fy, uy))

marks a significant change in the stiffness (inverse of compliance) of the device.

Comparatively low post-yield stiffness causes the elongation of the fundamental vibration

period of the structure and limits the force transferred through the isolator device to the

structure. In the frequency domain, the isolator device acts as a filter for high-frequency

components of the horizontal ground motion excitation. Furthermore, the hysteretic

energy dissipated by the isolation device during repeated cyclic horizontal motion serves

to reduce the lateral displacement of the isolator devices and to further reduce the inertial

forces transmitted to the isolated superstructure.

a. accelerations b. displacements

Fig. 3. Impact of period elongation obtained by seismic isolation on accelerations and

displacements of a structure.


Fig. 4. Typical relationship between displacement and force for a seismic isolation

device.


2.1.2. Types of Horizontal and Vertical Seismic Isolation Devices

Horizontal isolation systems are categorized as either elastomeric or sliding systems,

with different mechanical properties for each. These systems are briefly summarized in

the following paragraphs. Vertical isolation is also possible using spring-damper systems.

However, given the large mass of a typical nuclear reactor building, vertical isolation

would normally be provided only for individual, sensitive equipment, or for isolated

volumes such as a reactor control room, but not for the entire reactor building.

The first type of horizontal seismic isolation device is a laminated elastomeric

bearing, shown in Fig. 5. The elastomeric material in the bearings is natural rubber,

which has low horizontal shear stiffness but high vertical compressive stiffness. The low

horizontal stiffness and large deformation capacity of the elastomeric material is used to

provide the horizontal flexibility necessary for structural seismic isolation. The rubber is

present in horizontal layers that are bonded to horizontal steel shim plates in a process

called vulcanization, which forces the rubber to deform in shear and prevents vertical

buckling. The resulting bearing is more stable and its motion is more predictable. The

horizontal stiffness of the bearing depends primarily on the shear modulus of the rubber,

the bonded area of rubber, the total thickness of rubber and the lateral displacement of the

device. The reduced stiffness not only increases the building period, but also filters-out

higher order mode participation by capitalizing on the principle of modal orthogonality.

Additionally, elastomeric bearings re-center after earthquake shaking due to the forces

generated by elasticity of the rubber layers.

There are three main types of elastomeric bearings: low damping bearings, lead

rubber bearings, and high damping bearings. Two of the three, low-damping rubber

(LDR) and lead-rubber (LR) are suitable for nuclear power plant applications (Huang et

al. 2009).

Fig. 5 Cross section of a typical lead-rubber seismic base isolation bearing.


The second type of seismic isolation devices is the sliding bearing. Horizontal

flexibility of such isolation systems is provided by sliding, while their vertical stiffness is

provided by direct contact of the bearing elements. Control of motion of sliding bearings

is achieved by designing the shape of the contact surfaces, by designing the friction

coefficients between the contact surfaces, and/or by adding supplemental damping. The

friction specifically limits the amount of force transferred to the superstructure. Typical

European devices use friction and supplemental damping to dissipate energy. On the

other hand, the Friction Pendulum family of bearings that dominates the US market, shown in Fig. 6, uses a concave contact surface to control the motion of the bearing and

the isolated superstructure. As the isolated structure moves horizontally due to earthquake

excitation, the concave surface of the bearing forces it to also move upward, against

gravity, and slows it down. The radius of the concave contact surface, as well as the

friction coefficients between the sliding surfaces, are designed to give the Friction

Pendulum bearings desirable dynamic properties, such as stiffness and hysteretic

damping. More important, the concave surface insures re-centering, thus making the

Friction Pendulum bearing also suitable for use in nuclear power plant applications.

a. Single concave

b. Triple concave

Fig. 6 Friction PendulumTM

bearings (courtesy of Earthquake Protection Systems,

Inc.)

Vertical isolation can be achieved using spring and damper systems. Figure 7 shows a

typical example. Because of the stability issues associated with isolating large vertical

loads, development of vertical isolation systems for entire structures has lagged in

comparison to horizontal systems.

Outer

concave

plates

Inner

concave

plates Slider


Fig. 7 Helical springs and viscous dashpots (GERB) for vertical isolation.

2.1.3. Supplemental Damping and the Isolation Gap

Isolation systems are often designed with supplemental damping systems in order to

reach adequate levels of energy dissipation. Supplemental damping can also be used to

reduce high amplitude response, however, design of isolation systems should be

accompanied by the caveat that excessive amounts of damping can fundamentally inhibit

their efficacy.

Isolation gaps are designed to assure the superstructure responding to design-level

*/////ground motions remains isolated from the surrounding structure. However,

surrounding structure, or barrier, serves an essential role in reducing the beyond design-

basis displacements. The exact response and underlying mechanics of collisions with the

barrier should continue to be studied.

2.1.3. Seismic Isolation Design Basis

A base isolated nuclear facility structure has three components: the super-structure,

the isolation layer and the foundation embedded into the underlying soil, as illustrated in

Fig. 2. The seismic isolation devices are treated as structural components and thus fall

inside the domain of structural design code requirements. Thus, for this study the design

basis for all three components is consistent with ASCE Standards 4, 7 and 43. Beyond

these ASCE standards, NRC requirements must also be considered, as discussed in the

USNRC Standard Review Plan (NUREG 0800, Sect 2.5) and Regulatory Guide 1.122

Rev.1.

ASCE 43-05 defines a Design Basis Earthquake (DBE) using Probabilistic Seismic

Hazard Assessment (PSHA) to derive a Uniform (or Equal) Hazard Response Spectrum

(UHRS) for the site and modify it further using a Design Factor. The Design Factor is

calibrated to limit the annual seismic core damage frequency to under considering

the failure probabilities specified in current structural design codes and the design goals

proposed in ASCE 43.

The design goal is to reasonably achieve both of the following design objectives:

1. Less than 1% probability of unacceptable performance for the Design Basis Earthquake DBE ground motion defined in Section 2.0 of ASCE 43-05.


2. Less than 10% probability of unacceptable performance for a ground motion equal to 150% of the DBE ground motion defined in Section 2.0 of ASCE 43-

05.

A probabilistic design method proposed in ASCE 43-05 achieves these two objectives

simultaneously by specifying performance goals. The performance goal has a

quantitative and a qualitative part. Quantification of the performance goal is probability-

based: a mean annual hazard exceedance frequency (associated with the UHRS) and a

Target Probability Goal in terms of annual frequency of exceeding acceptable behavior,

, are defined. The pair defines the quantitative probabilistic performance goal,

expressed conveniently as a probability ratio . The qualitative performance

goal is defined using the Limit State concept to describe qualitatively the acceptable

structural behavior. ASCE 43 defines 4 Limit States, based on structural deformation

levels, to describe levels of acceptable structural damage. These damage states range

from significant damage with a structure close to collapse (state A) to no significant

damage with a structure in operational condition (state D). The qualitative and

quantitative portions of the performance goal are combined in the definition of the Target

Probability Goal . .

ASCE 43 defines 5 Seismic Design Categories (SDC). Categories 3, 4 and 5 are

associated with nuclear facility structures, systems and components (SSC) through

ANSI/ANS 2.26. Table 1 lists the Target Probability Goals for each of these categories.

Table 1. Target annual probability goals for ASCE 43 Seismic Design

Categories 3, 4, and 5 (from ASCE 2006)

SDC3 SDC4 SDC5

The target performance objective for nuclear power plant structures is SDC-5D. This

performance objective is associated with an annual probability of unacceptable

performance of 51 10 . Acceptable performance in fixed-base nuclear power plant

structures is described as essentially elastic behavior. This performance description is associated with conventional structural framing such as concrete shear walls and moment

frames and the foundations.

In base-isolated nuclear structures, the accelerations and deformations in structures,

systems and components are expected to be relatively small, such that the SSCs are expected to remain elastic for both DBE shaking and beyond design basis shaking. As

such, unacceptable performance of an isolated nuclear structure will involve either the

failure of isolation bearings or impact of the isolated superstructure against the

surrounding building or geotechnical structures. Therefore, unacceptable performance of


seismic isolation devices is defined as: 1) permanent damage to the isolation device, such

as tearing or disassembly; 2) exceedance of displacement limits of the device. The

acceptance criteria for isolation devices should be set such that there is a high confidence

of low probability of failure and that the overall target annual seismic core damage

frequency does not exceed the value required by USNRC. The superstructure and the

foundation of the isolated structural system are required to comply with the SDC-5D

performance objective set in ASCE-43.

2.1.4 Benefits and Challenges of Seismic Isolation

The potential benefits of seismic isolation to the nuclear power industry are

numerous. Seismic isolation can simplify the design, facilitate the standardization, and

reduce the cost to build new nuclear power plants, as well as improve the seismic

performance, enhance the safety margins, and enable a more accurate evaluation of the

probability of failure. The notion of seismic base isolation can be expanded to a general

concept of structural response modification and control. Thus, seismic isolation devices

can be viewed as a part of a portfolio of force limiting, vibration filtering, damping and

energy dissipation devices that can be passive or actively controlled. Response

modification technologies are being developed for conventional structural systems, but

have not yet matured to the level of application in construction achieved by the seismic

base isolation technology.

There are challenges for design, implementation and construction of base isolated

nuclear power plant structures. The first group of challenges is associated with the site of

the power plant. The seismicity of the site may include hazards stemming from near-by

faults, expected to cause large unit-directional velocity pulses and permanent ground

displacement at the site, as well as hazards from far-away faults capable of developing

extremely large ground motions, expected to induce long-duration ground shaking with

the majority of the seismic energy in the long-period (low-frequency) range. A base

isolated structure may suffer large single-pulse displacements or resonate with the ground

motions in the long-period range in these cases, presenting a design challenge.

The soil conditions at the site may present a design and an analysis challenge. Soil-

structure interaction (SSI) must be analyzed in design. If the soil column under the

foundation is soft, it may adversely alter the frequency content of the incoming ground

motion such that the energy is shifted towards the long-period (low-frequency) range.

Similar to the long-duration ground motion, this may cause the base-isolated structure to

resonate with the motion and develop large displacements in the isolation system. The

non-linear response of the isolators (and the underlying soil) requires time-domain non-

linear modeling and analysis capabilities. While such software exists, it has not been used

to design nuclear power plants in the past. Deployment of such time-domain analysis

method would be a departure from the frequency-domain analyses done to date, but may

be unavoidable because of the expected non-linear behavior of the seismic isolation

devices.

The second group of challenges is associated with the response of the seismic

isolation devices to ground motion excitation. In response to the horizontal components

of the ground motion, the isolation devices will sustain large displacements, requiring the


superstructure to move. If the superstructure is embedded and/or enclosed by the external

event shell, the seismic isolation gap needs to be large enough such that there is a very

small chance of the superstructure impacting the surrounding structures. Such impact

may damage the isolators and induce high-frequency high-intensity forces in the isolated

structure: thus, it should be avoided by conservative design, as well as the potential

provision of dampers or bumpers. In response to the vertical component of the ground

motion, the two types of isolation devices considered for nuclear power plant applications

(lead-rubber bearings and Friction Pendulum bearings) will transmit this excitation to the

superstructure without any reduction. Furthermore, overturning moments and, possibly,

uplift, may cause additional amplification of the vertical excitation. Additional research is

needed to develop a better understanding of the propagation of vertical ground motion

through soil and the seismic isolation layer into the structures. For equipment mounted

on slabs and for horizontal runs of piping, vertical accelerations may be as severe or more

so than horizontal motion.

The third group of challenges is associated with the transfer of the generated power

from the isolated part of the nuclear power plant, the citadel, to the balance of the plant.

Regardless of the medium used to transfer this generated power (electricity or heated

fluid), these conduits as well as many others (utilities, power, instrumentation) will have

to traverse the seismic isolation gap and accommodate the differential motion between

the two anchor points. Design of such umbilicals presents a mechanical engineering challenge. This is particularly the case for high-temperature, high-pressure fluids like

steam, where beyond-design-basis displacements of main steam lines may result in a

main steam line break (MSLB). Because the break would occur at the base isolation gap,

which is outside the reactor containment, valves would be available to isolate the steam

generator from the break, but the MSLB would still result in a severe thermal transient

for the steam generator and would disable its capacity to act as a long-term heat sink.

Conversely, Gen IV modular reactors that use compact, closed gas cycles (supercritical

CO2 or multi-reheat helium Brayton cycles) may mount the reactor and power conversion

system in a common base-isolated structure and eliminate the need for high-temperature

umbilicals. In addition, due to the very large mass of reactor buildings the design of

umbilicals must consider long-term soil settlement that can generate similar differential

offsets between the building and surrounding structures.

While serious, the three groups of challenges for use of seismic base isolation in

nuclear power plant structures have been encountered and successfully solved to

seismically isolated conventional, but critical, building structures such as hospitals and

lifeline facilities such as long-span bridges and LNG storage tanks.

2.2 External Event Shielding

Malicious crashing of a commercial aircraft creates loads that greatly exceed those

that might be expected from missiles generated by natural events such as tornados.

Aircraft crash into a structure can be expected to generate both significant global

structural demands that may exceed design basis earthquake or wind loads as well as

localized inelastic structural response in the vicinity of the impact point. Additional

demands on the NPP structures, systems and components in this beyond design basis


event may come from blast air pressure waves, penetrating aircraft parts and/or structural

debris acting as interior missiles, and fire. This section presents a comparative analysis of

structural layout concepts for an external event shell of a seismically isolated nuclear

power plant. Section 2.3 provides a comparison of the conventional reinforced concrete

construction and newly proposed steel-plate/concrete modular construction technologies

with respect to local behavior at the impact point. Section 6.4 presents design and impact

load response analysis for a sample conventional reinforced concrete external event shell.

In general, three major design variants can be considered for external event isolation

of a base-isolated reactor, as illustrated in Fig. 8. First, the entire reactor building

structure, including the external event shell, can be base isolated (Fig. 8A). This option

may be acceptable for very heavy structures, such as large LWR reactor buildings.

Second, base-isolated citadel may be placed inside a separate, decoupled external event

shell (Fig. 8B). Various options may exist to use the volume between the external event

shell and citadel productively, for example to house non-safety related equipment or

redundant safety related equipment, or to act as a duct for intake of external air for

emergency decay heat removal. Finally, the entire reactor building may be base isolated

and constructed fully below grade, with or without berms (Fig. 8C). These design option

involve issues related to construction cost and schedule, access for personnel and

equipment for operations and maintenance, umbilicals arrangements for transferring the

generated power from the plant and transmitting control commands to the plant, and

ducting arrangement if external air is used as a heat sink for decay heat removal. Also,

for partially or fully below-grade event shells, steady-state and seismic soil and

hydrostatic loads may be substantial and must be considered in design, and the pits must

have drainage and dewatering system.

The following sections describe each major design option in greater detail.

2.3.1 Coupled above-grade event shell and citadel

Conceptually, the simplest way to implement seismic base isolation is to isolate the

entire reactor building, including the external event shell, on a single base isolated

foundation, as shown in Fig. 7A. This approach allows conventional reactor building

arrangements to be adopted with minimal changes, and is the approach used in the

original ALMR design (ALMR 1994). Note, it is assumed that the foundations for

conventional reactors will be at least partially excavated to remove the shallow surface

soil and reach rock and/or firm soil layers. Thus, the seismic isolation gap required for

base isolated designs would have to be provided by the increased size of the excavated

opening: this configuration is commonly called the isolation moat.


Fig. 8 Three major external event isolation design options for base isolated reactors:

(A) above-grade, coupled event shell and citadel, on a single base isolated mat,

(B) above-grade, decoupled event shell and citadel, and (C) fully below-grade

construction.

The major design issue with a coupled event shell and citadel relates to the

accelerations that are imparted into the citadel during an airplane crash. These

accelerations are dependent upon the mass of the isolated structure, as well as the mass

and speed of the aircraft and its location and direction of impact. Increasing the isolated

mass is beneficial. For example, two or more reactor modules might be constructed on a

single base-isolated foundation. Likewise, the plant power conversion system might be

coupled to the reactor building foundation. This particular option has the additional

benefit of eliminating the need for high-temperature intermediate coolant pipes to span a

base-isolation gap to the power conversion system and to be designed to sustain

substantial deformations during earthquakes (discussed in Chapter 5).

Typical modular reactor designs may have reactor building masses ranging from

20,000 to 60,000 metric tons, with the larger values being possible when the power

conversion system is isolated on a single foundation along with its power conversion

system, or if two modules are isolated on a common foundation. Modern LWR reactor

buildings can be much heavier, as for the EPR and ABWR shown in Table 2. With the

trend toward passive safety systems for LWRs, reactor building sizes and masses will

likely decrease toward lower values, as for the early design of the ESBWR reactor

building given in Table 1.

A reasonable criterion to establish the minimum acceptable reactor building mass is

that the acceleration induced by aircraft crash be enveloped by the accelerations

originating due to seismic motion, so that the aircraft crash beyond design basis case does

not end up affecting equipment seismic qualification requirements. As discussed in

Chapter 6, for reactor buildings in the mass range of modular reactors, accelerations

induced by aircraft crash are likely to substantially exceed those induced by earthquakes.

Furthermore, it can be assumed that the design-basis aircraft crash may evolve over time

to include crashes of increasingly heavier aircraft.

Some possibility exists that advanced isolation devices could be designed that would

lock-out under aircraft crash conditions, transferring the crash loads to the ground and


limiting acceleration of the base isolated mass. Absent such an approach, however, this

report recommends that the external event shell for modular reactors be decoupled to the

base isolated mass of the citadel.

Table 2 Approximate total mass of typical LWR reactor buildings (Peterson et al.,

2005).

Reactor Concrete (m3)

Steel and Iron

(metric tons)

Total Weight

(metric tons)

1970's PWR (1000 MW) 22,600 7,500 59,500

ABWR (1350 MW) 67,500 18,500 174,000

EPR (1600 MW) 61,900 18,500 161,000

ESBWR (1500 MW) 29,200 8,900 76,000

The ESBWR underwent a substantial structural redesign that increased its weight from the value shown here.

2.3.2 Decoupled above-grade event shell and citadel

With a decoupled box-in-box eternal event shell, the shell transfers aircraft impact loads directly to its foundation. The event shell foundation can be coupled to the

foundation of the base-isolated citadel, such that the impact load is transfers the loads to

the ground, as it would be in a conventional, fixed-base NPP. The reactor citadel, inside

the event shell, is base isolated. A gap is provided between the event shell and the

citadel. This gap must accommodate the differential motion between the shell and the

citadel due to earthquakes (the seismic isolation gap) and that due to the deformation of

the event shell or its foundation caused by aircraft crash. Likewise, the gap must be

sufficient to accommodate rocking and torsion generated by eccentric loads, particularly

for hard rock sites.

Two basic geometries can be considered for external event shells. The curved surface

of a cylindrical event shell (Fig. 9A) increases the structural strength of the shell through

arching action. This allows impact loads to be transferred to the ground in part as tensile

and compressive stresses in the wall, rather than as bending stresses (except locally at the

point of impact). The seismic separation gap between the cylindrical shell and the base

isolated citadel can be smaller compared to that required for rectangular geometries

because radial symmetry makes it easier to accommodate maximum ground motion

demands and torsion caused by mass eccentricity, rocking, and vertical interactions.

Cylindrical event shells are particularly well suited for reactors that use large, dry

containments. In the AP-1000, this gap is used to duct external air flow in through vents

located near the top of the shell, down an annulus formed by a baffle plate, and then up

over the metal containment vessel to remove decay heat during accidents.

A rectangular external event shell has large, flat walls (Fig. 9B), which may enable

simpler construction methods. These walls must act as plates to transfer impact loads to

the side walls and roof diaphragms by bending, which then transfer the loads by bending

and shear to the base mat. Bending resistance of the exterior event shell wall may be


efficiently increased by integrating beam and columns into the inside surface of the wall,

either in a honeycomb or ribbed configuration. In particular, floors inside the event shell

can act as beams (Fig 9B) offering the required structural depth. This approach may be

attractive because this creates an accessible volume that is available to house equipment,

piping, cabling and ventilation ducting and to provide personnel access. However, if the

internal citadel is base isolated, a gap must be engineered into the floors, with a grating

and an expansion joint system, to allow for large differential motion between the shell

and citadel. The grating system must also function as a water-proof and debris-clean

barrier. Sumps and pumps are needed at the bottom of the pit to keep it dry, and must be

redundant in case of a system failure. Note that the equipment and piping that are

supported by the external event shell must be qualified for seismic loading expected in

fixed-base structures, while equipment which is mounted on or cantilevered from the

base isolated citadel can be qualified based upon the seismic response of the attachment

points of the seismically isolated structure.


(A) (B)

Fig. 9 Examples of (A) a cylindrical external event shell (AP-1000) and (B) a

rectangular external event shell (ESBWR).

The principal design challenge in decoupled event shell and citadel configuration are

the overhead cranes mounted on rails resting on opposite walls of the structure. Crane

rails can be supported on columns framed into the base isolated citadel structure, so that

the crane is also isolated from seismic motion, and so that differential motion between the

crane and reactor structures is minimized during earthquakes. But it is optimal to design

cranes to be capable of picking up loads close to walls: in a decoupled design significant

portions of the floor space attached to the fixed-base event shell structure may not be

reachable by the crane mounted on the base isolated citadel. This requirement may limit

the space for and configuration of the stiffening ribs or honeycomb on interior of the

external event shells, if rail-based reactor-service cranes are located in an open volume

above the reactor citadel (as is common practice for boiling water reactors, Fig. 9B).


2.3.3 Berms, below grade, and underground construction

Construction of a reactor below-grade requires that the reactor be constructed in a

deeply excavated space. Excavation can be performed by either open-cut or top-down

methods, with crane access being easiest with top-down methods. An important

constraint on below-grade construction is the fact that reactor buildings are commonly

very tall structures, with typical LWR, SFR, and MHR buildings being 70 to 80 m in

height from the bottom of the base mat to the top of the reactor building. The depth of

such excavations requires stabilization of the sides of the excavated opening against both

passive and active earth and water pressures and may require dewatering and water-

proofing. Furthermore, the substantial depth complicates the construction process and

increases its cost and schedule. If needed for plant operation, dewatering systems must

be redundant and seismic category SI.

The depth and cost of excavation can be reduced compacting the excavated soil into

berms around the periphery of the isolation moat, at the expense of increasing the

footprint occupied by the reactor substantially and potentially complicating the capability

to locate supporting structures such as turbine buildings. In some reactor designs the

reactor floor is at the external grade level, such that the structures above grade support

and enclose the reactor high bay space. If the reactor refueling floor is sufficiently

hardened, it is possible that the structure enclosing the high bay space above the floor

does not need to be engineered to have the strength required to exclude penetration by

aircraft. However, this may require complex engineering and security analysis for

activities such as refueling where the crane is being used to access equipment below the

reactor floor, and to analyze for effects large fires on the refueling floor. If the above

grade crane-bay structure is hardened to exclude aircraft, then the building falls into the

above-grade category.

For reactors that use external ambient air for passive decay heat removal, as is the

case for most SFR designs, the air intake and exhaust vents are normally located

substantially above grade to prevent snow drifts or drifting sand from blocking the intake

vents, and to reduce accessibility for potential sabotage. If below grade construction or

berms are used, the associated chimneys extending above grade must be hardened

sufficiently to survive aircraft impact loads.

Underground siting has also been studied for nuclear power plants. The California

Energy Commission (CEC, 1978) considered two major options, (1) mined-cavern siting,

and (2) berm containment and cut-and-cover construction. In mined cavern siting, large

caverns, some 30-m wide, 60-m high, and a few hundred meters in length, would be

excavated from solid rock. The nuclear reactor is then constructed in the cavern, after all

rock motion has ceased. In the second option, a large pit approximately 120-m in

diameter and 45 m deep is excavated, and the nuclear power plant built in the pit. The

excavated material is then mounded over the entire plant. The CEC study identified

several potential benefits from underground siting, including more effective containment

of radionuclides following a severe core damage accident, improved earthquake

resistance, easier decommissioning, improved resistance to sabotage, and urban siting

(CEC, 1978).


The CEC study concluded that construction of both mined-cavern and berm contained

plants would be technologically feasible. Construction costs for n-th of a kind plants

were expected to be 14% higher for berm-contained plants, and 25% higher for mined-

cavern plants, compared to a surface-sited plant. Likewise, construction time for n-th of

a kind plants was expected to be 22 months longer for berm-contained plants, and 19

months longer for mined-cavern plants. While the study found potential benefits, it

concluded that underground siting should not be mandated, due to uncertainty over costs

and construction time, the existence of what appear to be moderately effective and less expensive technical alternatives, and the opportunity to implement remote siting. Today, because reactor structures can be engineered to accommodate the effects of severe

external events, the additional cost and schedule delays associated with underground

siting are likely still not warranted.

2.3 Steel-Plate/Concrete Modular Construction

Conventional, fixed-base, nuclear power plant structures are constructed using the

reinforced concrete structural shear wall elements to form the primary gravity and lateral

load structural systems. In addition to their load-carrying ability, such walls offer good

radiation protection, provide a satisfactory pressure barrier, and offer good fire resistance.

The isolated superstructure and the foundations of seismically isolated nuclear power

plants may also be constructed using this conventional technology.

The structural reinforced concrete shear walls are erected using the cast-in-place

construction method. The forms for such walls are erected first. The reinforcement is

placed into the forms, usually using pre-assembled reinforcement cages. Finally, the

concrete is poured into the forms, vibrated into place, and left to cure to a strength that

allows for removal of forms. While there is ample experience with such construction

method in the industry, three obstacles still remain. One is the design of the forms to

sustain the weight of the flowing concrete: this limits the height of the pouring lifts, thus

constraining the speed of construction. Second is the time needed for concrete to cure to

the level where forms can be removed to advance construction. Use of slip-forming for

cylindrical walls can accelerate construction. And third is the need to completely remove

the forms such that, over time, wood debris left behind does not cause unwanted

corrosion. Liner corrosion problems have been observed at existing U.S. nuclear plants

due to wood debris left during construction.

The solution for these construction problems is to use so called stay-in-place forms. Such forms are, by themselves, structural elements that are capable of sustaining the load

of flowing concrete and additional formwork above the pour level. In addition, they can

be composited with the poured concrete to become part of the larger structural element.

Stay-in-place forms can, furthermore, be pre-fabricated and pre-assembled into large

modules, enabling significant speedup in construction.

Stay-in-place forms have been made using a variety of materials. It is not uncommon

to use thin reinforced concrete plate elements, or plates made of fiber-reinforced mortars

or polymer composites to form structural walls and leave them in place. However, the

most common material used for stay-in-place forms is steel plate. As shown in Figure

10(A), steel plates forms can be pre-fabricated as modules of significant size, lifted and


assembled into place, and welded at the contacting seams before additional reinforcement

is placed into them (if needed). The concrete pour is then the same as in conventionally-

formed walls. Supplementary steel reinforcement may be required in thicker stay-in-place

form walls and on the top side of stay-in-place form slabs where no steel plate exists;

however, the amount of rebar is greatly reduced. The steel plates are advantageous in that

they do not provide unidirectional resistance like rebar. Thus, material savings of

concrete, forms, and steel bars/plates were estimated to be 6%, 97%, and 19%,

respectively (Takeuchi, 1995). Accordingly, a shift towards prefabrication and on-site

skilled steel workers (i.e. welders) is anticipated.

The principal advantage of steel stay-in-place forms is that the form steel can be

composited with the interior concrete to form a composite steel-concrete structural

element (sometimes called sandwich walls). Steel-concrete composite construction is

regulated by structural design code documents: ASCE 7-05 for design of composite

structural systems, and ANSI/AISC 360-05 and ANSI/AISC 341-05 for design of

composite structural elements for non-seismic and seismic applications, respectively. In

particular, a number of research studies support the code design provisions for composite

steel plate shear walls, where structural concrete is encased by steel plates and the bond

between steel and concrete is maintained to ensure composite load carrying action. The

additional confinement of concrete and inhibition of concrete spalling by the surface

plates enable higher ultimate strengths and increased ductility of the composite members,

particularly under cyclic loading conditions (Munakata, 2009). Figure 10(B) shows the

tie bars and steel studs used to maintain spacing of plates during concrete pouring as well

as achieve the composite action between steel and concrete. In addition, steel members

known as ribs or stiffeners are continuously connected to the steel plates such that they

can sustain the hydrostatic pressure of flowing concrete and maintain their form.

An additional advantage of stay-in-place steel forms is the ability to pre-fabricate

large and complex modules. Using ship-building techniques, modules comprising a

variety of structural openings, corners and attachments can be pre-fabricated in a

controlled factory environment and shipped to the construction site. The only limit is the

shipping and lift-and-place capability. Figure 11 shows a recent installation of a large

(over 700 ton) steel module assembled from factory-prefabricated modules at an AP-

1000 NPP construction site in Sanmen, China. Such modules can be pre inspected at the

construction yard and match-assembled to ensure easy installation and welding on site (a

technique long-used in long-span bridge construction).

Earlier experience with steel-plate composite structure construction exists from

application by General Electric to boiling water reactor containments. The GE Mark III

fleet (STRIDE Program) has a drywell vent structure designed and built as a composite

steel-plate/concrete structure. The Mark III drywell vent structure is about 46 m (150 ft)

diameter and about 18 m (60 ft) high with conventional concrete construction above it,

above the suppression pool. The Mark III reactor pressure vessel pedestal and shield wall

use the same technology but smaller diameters. These Mark III's have operated

successfully for over 20 years at sites including River Bend, Perry, and Grand Gulf

(Solorzano, 2009).


The GE Mark III pedestal and shield wall fabrication used modules assembled in a

fabrication shop area. A special on-site shed was used for fabrication because the

modules were so large, circumferentially and vertically. Workers fitted the modules up,

placed alignment devices and match-marked for re-assembly at the site. The people who

were responsible for assembling the modules on site were present to witness this fit-up

operation.

Steel-plate/concrete construction has been identified as a very promising technology

for construction of new nuclear power plants (Schlaseman, 2004). However, while earlier

experience exists with BWRs, this technology still entails significant challenges.

The first group of challenges concerns the design and modeling of steel-

plate/concrete walls. The experimental data available for such walls has been obtained by

testing scaled-down specimens, with overall thickness between 100mm and 300mm and

steel plate thickness between 2mm and 6mm. The prototype wall thicknesses are larger

than 1,000mm while the plate thicknesses remain between 6mm and 12mm. Recognizing

that full-scale test may not be possible using the existing structural laboratories, more

tests on correctly scaled specimens are needed to establish the benchmarks for calibrating

computer models for thick steel-plate/concrete walls. That said, the existing computer

models for composite steel concrete structures need to be validated for situations where

structural shapes are used to tie the two plates together, to carry the bi-axial shears, and

where studs are provided to prevent the thin liner plates from debonding during pouring

or unexpected internal out-gassing pressures. For limitation of buckling the studs also

play a useful role because only a small amount of support is needed to prevent buckling

from occuring.

The second group of challenges concerns construction of steel/plate composite walls.

While construction of the wall itself is made much easier, implementation of the wall-to-

floor connections is challenging. The GE Mark III pedestal, for example, required large

embedded plates in the base mat foundation, with alignment bars and backup blocks for

welding. Furthermore, design and implementation of structural connections between

steel-plate/concrete and conventional reinforced concrete elements has not been

extensively studied, although again the top of the GE Mark III containment composite

structure is designed to provide a transition to conventional reinforced concrete

construction. Since the concrete curing process emits a large amount of water vapor

which cannot permeate the steel layer covering the concrete, vents are needed in stay-in-

place form structures. As walls get thicker, elaborate venting/cooling system become an

important design issue to remove large amounts of water that remain deep within

concrete members as a byproduct of curing. If not removed from the structure, small

water deposits have the potential to initiate and propagate cracks under high

temperatures, such as in response to fires. Finally, implementation and anchoring of

piping and equipment attachments to the steel skin of steel-plate/concrete walls remains

an important design issue.

The third group of challenges concerns inspection and service of steel/plate

composite walls over the life of the nuclear power plant. It is universally recognized that

the composite walls offer superior resistance to extreme loading conditions, such as

aircraft impact and earthquake loading. Under impact, the steel skin increases the


resistance of the wall to local perforation, scabbing and penetration (shear cone

punching) compared to conventional reinforced concrete walls. Under earthquake

loading, properly designed steel-plate/concrete walls (with details to delay elephant-foot

buckling and formation of buckled compression diagonals) should be significantly more

ductile than the correspondingly reinforced conventional reinforced concrete walls.

However, long-term degradation of steel-plate/concrete walls is not fully understood,

although experience exists with steel liners commonly used on the internal surfaces of

reactor containments, such as the GE Mark III containment. The exposed steel plates

may need to be protected against corrosion. Inspection methods to establish the

conditions of inaccessible portions of the walls have not been evaluated, particularly

concerning the state of bond between steel plates and concrete. Behavior of such walls

under elevated temperatures during accidents and fires has not been studied at the size

and scale used in nuclear power plants.

Finally, the forth group of challenges concerns decommissioning of nuclear power

plants after the end of their service life. Methods for effective deconstruction of the steel-

plate/concrete walls have not been investigated. In particular, it is not clear how to deal

with portions of such walls that may be irradiated or otherwise contaminated during the

regular service life of a nuclear power plant.


(A)

(C)

(B)

Fig. 10 Schematic illustration of steel plate concrete construction method (A, B) and a

typical Westinghouse AP-1000 CA01 structural sub-module (C) (credit Westinghouse).


Fig. 11 An example of steel-plate concrete construction, a ~20-m high AP-1000

auxiliary building is assembled from factory-prefabricated modules and is set in

place by a heavy lift crane in Sanmen, China (Credit Westinghouse).

2.4 References

ALMR Reactor Facility Seismic Analysis, Bechtel National, Inc., October 1994.

California Energy Commission, Underground Siting of Nuclear Power Reactors: An Option For California, Staff Report, Nuclear Assessments Office, June, 1978.

Chopra, Anil K. Dynamics of Structures. New Jersey: Pearson Prentice Hall, 2007.

Huang, Y.-N., Whittaker, A. S., Kennedy, R. P. and Mayes, R. L., "Assessment of base-

isolated nuclear structures for design- and beyond-design basis earthquake shaking."

MCEER-09-0008, Multidisciplinary Center for Earthquake Engineering Research,

State University of New York, Buffalo, NY, August 2009

Malushte, Sanjeev R., and Andrew S. Whittaker. Survey of Past Base Isolation Applications in Nuclear Power Plants and Challenges to Industry/Regulatory

Acceptance. 18th International Conference on Structural Mechanics in Reactor Technology (SMiRT 18, 2005). K-10-7.

Munakata, Yoshinari, Takashi Maki, Yoshiyuki Sato, Keiji Sekine,Takamasa Nishioka,

Nobuyuki Niwa, and Osamu Kontani. Study on Radiation Shielding Performance of Reinforced Concrete Wall: Loading Test on Concrete Walls and Modeling of

Concrete Cracks. 20th International Conference on Structural Mechanics in Reactor Technology (SMiRT 20 - 1866, 2009).


Peterson, P.F., Haihua Zhao, and Robert Petroski, Metal And Concrete Inputs For Several Nuclear Power Plants, Report UCBTH-05-001, UC Berkeley, February 4, 2005.

Schlaseman, C. Application of Advanced Construction Technologies to Nuclear Power Plants, Technical Report, MPR-2610, Revision 2, P. 132, US Department of Energy, Sept. 24, 2009.

Solorzano, E., Personal communication, October 20, 2009.

Takeuchi, M., H. Akiyama, M. Narikawa, K. Hara, H. Tsubota, and I. Matsuo. Study on a concrete filled steel structure for nuclear power plants: (Part 1) Outline of the

structure and the mock-up test. 13th International Conference on Structural Mechanics in Reactor Technology (SMiRT 13 H02 1, 1995).

US Nuclear Regulatory Commission (NRC) Regulatory Guide 1.165 (1997),

Identification and Characterization of Seismic Sources and Determination of Safe Shutdown Earthquake Ground Motion


3.0 SIMULATION VERIFICATION AND VALIDATION ISSUES

To use simulation results in USNRC Design Certification of a base-isolated nuclear

reactor, both the structural engineers and the regulators will have to have a sufficient

level of confidence that the results provide an accurate representation of the physical

system. The process of establishing this confidence in commonly referred to as

verification and validation (V&V). This process, which can often be very intensive,

includes establishing both that one has correctly solved the desired system of equations

(verification) and that the chosen model sufficiently represents physical reality for the

intended purposes (validation).

This chapter reviews some general challenges for the modeling and simulation of

base-isolated structures based on limited historical experiences, currently accepted

simulation methods, and the unique validation requirements to represent the non-linear

behavior of base isolators. Similar V&V issues exist for the modeling of structural

response to large aircraft crashes, particularly the inevitable localized inelastic response

of the structure. However, because this report specifically recommends that base isolated

structure be decoupled from the external event shells, specific V&V issues for aircraft

crash modeling are not discussed here.

3.1 Historical Experience with Seismic Base Isolation

The response of base-isolated structures to historical earthquakes will be an important

source of information for the validation of computational models used for future designs.

Currently this data set remains quite limited due to the small number of existing isolated

structures and the relatively long earthquake return period (i.e. few large earthquakes).

As more isolated structures are completed and instrumented, additional data should

become available over time that will be invaluable in the assessment of computational

modeling tools.

The best source of historical data for base-isolated structures in the US is the magnitude

6.8 Northridge earthquake that struck the Los Angeles area on January 17, 1994. Three

base-isolated, steel-frame structures experienced strong ground motion recordings above

0.20 g. Table 3 summarizes the data for these three structures. While the USC Teaching

Hospital appears to have behaved well, it is apparent from these results that not all

isolation systems behaved as expected since the two other structures experienced roof

accelerations larger than the peak ground acceleration. This amplification was not

expected for these isolated structures, and may have been caused in part by retrofits that

compromised the isolation gap. Although the peak accelerations exceeded the PGA, it is

worth noting that these accelerations are likely less than what the non-isolated peak

accelerations would have been assuming the superstructure had a small non-isolated

period and was located at a site with adequately-strong soil composition. These concepts

will be developed further in Section 6.3.


Name Epicenter Distance

(km)

Isolation System

Peak Horizontal

Ground Acceleration

(g)

Peak Horizontal

Roof Acceleration

(g) University of Southern California Teaching Hospital

36 Lead-Rubber

Bearings 0.37 0.21

L.A. Fire Command and Control Center

38 High Damping

Rubber Bearings

0.22 0.32

Santa Monica Private Residences

21 GERB Spring-

Damper 0.44 0.63

Table 3. Summary of response for base-isolated structures with strong ground motion

from the Northridge Earthquake (Clark et al 1996).

The Teaching Hospital at the University of Southern California experienced large

ground accelerations in the Northridge earthquake and demonstrated the effective

reduction of accelerations within the isolated structure. The structure is an eight-story

steel frame supported by 68 lead-rubber isolator and 81 elastomeric isolators (Clark et al

1996). The periods for the first mode of the structure are 1.32 in the east-west direction

and 1.38 in the north-south direction, compared to 0.92 and 0.76, respectively, if the

structure had a fixed base (Nagarajaiah and Xiaohon 2000). Significantly, the peak roof

acceleration was observed to be 43% less than the peak ground acceleration.

The analysis of the USC hospital response to the Northridge quake is supported by

the characterization of the bearings through cyclic loading tests performed by Dynamic

Isolation Systems. Nagarajaiah and Xiaohon (2000) used this data to develop bearing

properties for a smooth bilinear hysteretic model and a simplified equivalent linear

displacement model. Their analysis using the computer code 3D-BASIS showed that the

nonlinear hysteretic model more closely matched the observed response spectra than the

equivalent linear-isolation system. Qualitative observation of the results shows that the

linear models tend to under-predict the structural accelerations. However, these

discrepancies are not quantified in the analysis by Nagarajaian and Xiaonon.

The experience from the USC Teaching Hospital demonstrates the potential

effectiveness of base-isolation and the ability to characterize the bearings and model the

system response, as well as the difficulty in selecting appropriate models to describe the

system accurately. For the simulation of base-isolated nuclear facilities, it will be more

difficult establish a case for the models used in the analysis since the goal of such efforts

are predictive and real data from an isolated reactor in a strong earthquake will probably

not be available for some time.

The ability to pre-predict the response of the USC hospital (i.e. accurately predict the

response before evaluating the observed response) could be one important validation

metric for analytic methods used in isolated reactors. Since nuclear structures are

generally stiffer than the USC hospital, their dynamic response would more closely


approximate the ideal isolated rigid structure (Chopra 2007). The use of this historical

case would therefore be useful for validation since the USC hospital structural response

may be more difficult to characterize than nuclear structures and will help to establish

confidence in the modeler and the model approximations used for the response of lead-

rubber bearing isolators.

The Los Angeles County Fire Command and Control Facility (FCCF) is a two-story

steel frame structure completed in 1990 with 32 high-damping rubber isolators

supporting a structure of 1930 metric tons (4230 kips). The isolation system is designed

for a peak displacement of 24.4 cm (9.6 in) and has the unusual feature of chains installed

at the center of each bearing that are set to engage at displacements of 31.8 cm (12.5 in)

(Clark et al. 1996).

The response of the FCCF included several high frequency spikes in the east-west

direction that resulted in amplification ratios much greater than 1.0. The behavior in the

north-south direction resembled the expected behavior with amplification ratios on the

order of 0.5. The divergent east-west response is believed to be the result of a

compromised isolation gap where sacrificial elements were re-enforced after being

damaged in the 1991 Sierra Madre earthquake. The contact between the isolated

structure and the re-enforced elements would lead to pounding consistent with the

observed high frequency spikes.

The experience with the FCCF shows that special care needs to be taken for any

maintenance and repair work that may impact the isolation gap for a base-isolated nuclear

structure. Any changes to structures or c

Documents

09-004 UCB Gen IV Base Isolation Rpt Final